High‐Temperature Oxidation‐Resistant Printed Copper Conductors

Advanced materials, electrically conductive and oxidation resistant, are frontrunners for technological advancements in cutting‐edge high‐temperature electronics. Rational design and manufacturing of hierarchical material structures is indispensable to achieve such disparate functionalities. Here, high‐temperature copper–graphene conductors, through additive manufacturing, which prohibits oxygen adsorbates and serves as the barrier for oxygen migration to enable electric stability and reliability at high temperatures, are reported. The combination of graphene and alumina surface passivation enables the electric stability of copper–graphene under thermal impact above 1000 °C. The findings shown here, the synergistic combination of high conductivity and oxidation resistance, enunciate the passivation capabilities for additively manufactured flexible electronics operating under harsh conditions.


Introduction
Copper is one of the cutting-edge materials for advanced electronics, however, it is susceptible to oxidation under hightemperature environments due to its relatively low oxidation potential, and this is prominently observed in nanostructured copper materials. [1,2] The formation of nonconductive oxide on the surface of copper creates an insulating effect between the nanostructures, which expedites the decrease in electrical conductivity. Improving the oxidation resistance of copper without hampering its electrical conductivity has been a field of interest www.advelectronicmat.de Developing such strategic conductor material can offer tremendous benefits, thus expanding the extreme environment electronics. [9] In order to enhance the stability of hybridized copper-graphene (Cu-G) conductors under harsh environments, [10] herein, we describe surface modification strategies using two dissimilar materials, namely graphene (Cu-G/G) and alumina (Cu-G/Al 2 O 3 ), for enabling the printed copper conductors to endure oxidation as a result of elevated temperatures. Coating dissimilar materials, carbon (graphene) or ceramic (alumina) on copper seems a deliberate method to enhance the oxidation resistance without affecting the electrical conductivity of copper. Graphene and alumina are known for their high thermal stability and oxidation resistance. [11][12][13][14] As the oxidation kinetics depend on the rate of the inward diffusion rate of oxygen, the presence of the suggested graphene and alumina surface modification provides electric stability up to 660 °C (a resistivity change from 0.004 to 0.01 Ω cm) and 590 °C (a resistivity change from 4.38 × 10 −5 to 3.09 × 10 −4 Ω cm), respectively. Briefly, the Cu-G/G and Cu-G/Al 2 O 3 materials display a substantial increase in oxidation resistance whilst maintaining the electrical conductivity of base copper nanostructures. The oxidation resistance can be further improved by combining a layer of graphene followed by alumina, enabling the sample to survive upwards of 950 °C (a resistivity change from 1 × 10 −5 to 8 × 10 −5 Ω cm). This study facilitates a method for developing novel copper conductors with superior oxidation resistance with high electrical conductivity, as demonstrated by subjecting a Cu-G/G sample to a hydrogen-flame torch test up to a temperature of 1000 °C with a change of 18.75% in the resistance and evaluating a Cu-G/Al 2 O 3 resistance-temperature detector (RTD) up to a temperature of 500 °C and 600 °C, respectively, with a sensitivity of 0.0316 Ω °C −1 and 0.0327 Ω °C −1 .

Results and Discussion
The base material for this study consists of a conductive ink from a mixture of copper (111) nanoplates and dopamine hydrochloride to form the hybridized Cu-G material, [15,16] resulting from an in-situ conversion of polydopamine, a potent source of graphene, when subjected to higher temperatures. From our previous works, the XPS data clearly confirms the presence and conversion of polydopamine to nitrogen-doped graphene. [15] Graphene is utilized to form the hybridized Cu-G materials and provide benefits like enhanced corrosion resistance and thermal stability, [10,17] thus resulting in an enhanced metallic conductor with an amalgamation of high electrical conductivity of copper and high thermal stability of graphene. Figure 1a shows the schematic of surface passivation where either graphene or alumina coatings in varying thicknesses can be deposited onto the printed Cu-G conductors. The passivated layer prevents the migration of oxygen molecules toward the Cu-G at high temperatures, thus improving its overall oxidation resistance ( Figure 1b). Figure 1c depicts high-temperature electric reliability testing setup for the passivated Cu-G features, while Figure 1d shows a summary plot of the alumina and graphene passivation techniques to improve the oxidation stability of Cu-G to 590 °C and 660 °C, respectively, with the initial resistivities being 4.12 × 10 −5 Ω cm for Cu-G, 4.04 × 10 −3 Ω cm for Cu-G/G and 4.38 × 10 −5 Ω cm for Cu-G/Al 2 O 3 . The variance in resistivities is due to the placement of electrode connections of the top-most layer of the sample, with the Cu-G/G sample being coated with a 20 µm graphene layer (essentially graphite) and as graphene deposited on top has lower conductivity compared to in-situ converted bare Cu-G layer, this www.advelectronicmat.de leads to a variation in the initial resistivities among the three samples. Figure 2a shows the plot depicting the resistivity versus temperature curves for varying graphene concentrations, while an optimum graphene concentration of 0.237 wt% shows the robust oxidation resistance of copper below 450 °C on flexible alumina ribbon ceramic substrates. The flexibility of the ceramic was established by bending tests with printed Cu-G features in one of our previous works. [18] It is essential to note that the resistance values differ from sample to sample due to printing and sintering conditions, formation of the percolation network, and impurities that may be introduced between manufacturing and testing process. The sample thickness was measured and can be observed in Figure S4, Supporting Information. Figure 2a inset shows the conductivity dependence of increasing graphene concentrations. It is evident that an increment in the graphene weight reduces the conductivity to a certain extent. Figure S1, Supporting Information, shows the temperature-dependent resistance (R-T) of graphene control experiment, with the starting resistance being ≈900 Ω, and hence copper is necessary to have a high electric conductivity (or low resistance). Figure S12, Supporting Information, shows some low-temperature measurements with residual resistance slightly changing after every cycle. Figure 2b shows the low leakage current of Cu-G at 9 nA, which is beneficial for a plethora of electronic applications. The sample consisted of four and three fingers on each side respectively with a spacing of 2 mm, and the test is conducted with a voltage of 100 V under ambient conditions. The finger length is 10 mm and the thickness is approximately 15 µm. Figure 2c shows the resistance stability of Cu-G under ambient conditions for a duration of 24 h, while Figure 2d shows the aging of the Cu-G at 300 °C for 55 h. Without surface passivation, the Cu-G is relatively stable but undergoes some degree of oxidation as a result of long-time exposure to the elevated temperature at 300 °C. SEM images of the Cu-G conductor before and after high-temperature testing can be seen in Figure S5, Supporting Information. Figure S5a, Supporting Information, shows the conductor before high-temperature testing, as all nanoplatelets can be seen. Figure S5b, Supporting Information, shows the oxidized conductor after high-temperature testing, indicated by the formation of a scale-like structure over the nanoplates. Figure S9, Supporting Information, shows the Raman spectra of the conductor before and after the test. Figure S10, Supporting Information, shows the SEM and EDS spectra of the conductor after the aging test and its oxidation when the temperature was ramped up to 400 °C. Figure 3 shows the graphene passivation effect on the electric stability of printed Cu-G/G at elevated temperatures. Figure 3a shows the schematic image of the graphene coating layer on the printed Cu-G conductor feature. Figure 3b Figure 3c shows the thicknessdependent resistivity-temperature characteristics of graphenecoated Cu-G, on which 20 µm graphene coating displays the highest stability at a temperature up to 660 °C (a resistivity change from 0.004 to 0.01 Ω cm), compared to its other counterparts. The primary failure mechanism would be the oxidation of the top surface of the graphene thin film, thus exposing the Cu-G conductor underneath, causing a breakdown of the device. Figure 3d shows the Raman spectroscopy of the graphene-passivated Cu-G before and after high-temperature evaluation. After the test, the graphene coating oxidizes, as it is in the range of graphene oxidation temperature, thus exposing the inner Cu-G to oxygen, causing its electric breakdown. Figure 3e shows the graphene-passivated Cu-G under the thermal impact of hydrogen-flame torch test. As the torch testing is dynamic and lasts for a short duration compared to the steady state thermal test, it survives to a temperature above 1000 °C, with the resistivity showing a downward trend from the initial, which can be attributed to the negative temperature coefficient of resistance of graphene. [19] The oxidation resistance of Cu-G/G can be further enhanced by using alumina coating (Figure 3f). The sample survives up to 950 °C, compared to 660 °C without alumina coating. Figure 3f inset shows the plot depicting the aging test of Cu-G passivated with a coating of graphene monolayer feedstock and an additional layer of alumina at 1000 °C for 70 h. When the Cu-G conductor material is exposed to high-temperature oxidative environments, the passive coating layer is ineffectual for longer exposures. [20][21][22][23] Figure S6, Supporting Information, shows the in situ XRD at specified temperatures, depicting the performance of Cu-G/G powdered sample. As can be inferred from the figure, graphene starts oxidizing around 600 °C. Therefore, increasing the thickness of alumina coating could provide the considerable benefits against oxidation at elevated temperatures as it could delay the inward oxygen diffusion. SEM images and EDS mapping of the oxidized conductor, indicated by the formation of scaly, bead-like structures on the copper plates has been depicted in Figures S3 and S7, Supporting Information. Figure 4a further shows the alumina passivation strategy of Cu-G (Cu-G/Al 2 O 3 ), while Figure 4b shows the planar view of the SEM image with the embedded elemental mapping to depict the alumina coating. Figure S2 further shows the electric stability of Cu-G/Al 2 O 3 under a corrosive environment (0.5 m H 2 SO 4 ) in comparison with the base Cu-G. Figure 4c shows the thickness-dependent R-T characteristics of Cu-G/ Al 2 O 3 . It should be noted that 100 nm alumina coating outperforms its counterparts and can survive well above 590 °C (its resistivity changes from 4.38 × 10 −5 to 3.09 × 10 −4 Ω cm), thus extending the life of base Cu-G by approximately 140 °C. The failure mechanism is most likely the delamination of the alumina thin film due to high temperatures, thermal www.advelectronicmat.de stresses, and temperature coefficient mismatching, causing the conductor underneath to be exposed and hence leading to a breakdown. Figure S8 shows the oxidized conductor after steady-state high-temperature testing. Figure S11 shows the cross-sectional SEM view that was utilized for calibrating the thickness for the ALD equipment. Figure 4d indicates the aging test of Cu-G/Al 2 O 3 at 500 °C for 90 h. The insets in Figure 4d shows the Cu-G/Al 2 O 3 materials at intervals of 50 h and 100 h, respectively, depicting its scattered oxidation. Figure 4e shows a resistance-temperature detector (RTD) of Cu-G/Al 2 O 3 up to a temperature of 500 °C. The sensitivity can be calculated by the product of the resistance at the reference temperature and the temperature of coefficient, α, determined by where R 1 and R 2 are the initial and final resistances, and T 1 and T 2 are the initial and final temperatures. The temperature coefficient value of the RTD using Cu-G/Al 2 O 3 comes out as 2.36 × 10 −3 °C −1 , and hence the sensitivity of the RTD sensor is 0.0316 Ω °C −1 . By means of a linear fit, the equation for temperature in terms of resistance is of the form T = m*R + C, where T is temperature and R is the resistance, and here m = 30.83 and C = −358.93. Figure 4f shows the RTD performance up to a temperature of 600 °C with an α value of 2.16 × 10 −3 °C −1 and a sensitivity of 0.0327 Ω °C −1 . Additionally, it shows m = 27.29 and C = −329.34. The inset of Figure 4f shows the comparable performance of the Cu-G/ Al 2 O 3 RTD with a commercial platinum thermocouple on a normalized scale (range of −50 °C to 200 °C limit), with an α value of 3.66 × 10 −3 °C −1 and a sensitivity of 0.40 Ω °C −1 . Figure S13, Supporting Information shows the cycling and sensitivity plots of the Cu-G/Al 2 O 3 conductors. There is a declining trend in the sensitivity, which is intuitive given the slow oxidation of the conductor after every ramp.
Several applications require wireless communication between two devices, for instance, it can be beneficial in a flight to have devices wirelessly communicate rather than using wires, which helps in reducing the overall payload of the flight. Figure 5 shows the Arduino setup for the RTD to enable wireless communication with the local area network. It consists of an ESP32 module, which is essentially an economical, lowpower system on a chip with built-in Wi-Fi and Bluetooth. The sensor is thus capable of transferring the resistance as well as the associated temperature to the local area network via WiFi connectivity. Figure 5a shows the image of the device configuration. Figure 5b depicts a plot with the measured resistance corresponding to the Cu-G sensor output temperature and a comparison with a commercial thermocouple. The fabricated RTD performs akin to the commercial thermocouple. This sensor displays a sensitivity of 0.01 Ω °C −1 and an α value of 0.291% °C −1 . [24] Figure 5c shows a snippet of the webpage rendering temperature updates in real time.

Conclusion
We present printed copper nanostructures to enable electric stability and reliability at elevated temperatures. The surface passivation of hybridized Cu-G conductor by printing graphene with a thickness of 20 µm shows the optimum electric stability at a temperature of 660 °C, while the alumina passivation with a thickness of 100 nm ensures www.advelectronicmat.de the copper-based conductor stable up to 590 °C. Additionally, combining the passivation methods enables the conductor to survive a temperature upwards of 950 °C. Evidently, the oxidation resistance of Cu-G is significantly improved after surface passivation, while the printed conductor can survive hydrogenflame torch testing to a temperature 1000 °C. This work provides a pathway for creating printed copper conductors with oxidation resistance at elevated temperatures. The utilization of graphene and alumina coatings on printed copper nanostructures is a step forward in the realization of materials with high oxidation resistance in combination with high electrical conductivity, which can be further explored for a wide range of high-temperature electronic applications.

Experimental Section
Copper-Graphene (Cu-G) Ink Preparation: Copper nanoplate feedstock was prepared to utilize in previous recipe. [15] To the obtained feedstock, 20 wt% hydroxypropyl methylcellulose solution (2 wt% in deionized water) was added, with additional 55 wt% water, resulting in an ink consisting of 25 wt% conductive nanostructures. Dopamine hydrochloride was added to the ink ranging from 0.3 to 1.5 wt%, and the ink was mixed in a Thinky Mixer (ARE-310) for 30 s at 2000 rpm to attain homogeneity. For determining the graphene concentration in the print after sintering, we assume 100% dopamine hydrochloride conversion to graphene. This would result in the formation of 0.237 wt% graphene from 0.3 wt% dopamine hydrochloride and similarly, 0.6, 0.9, 1.2 and 1.5 wt% dopamine hydrochloride will result in 0.474, 0.711, 0.948 and 1.185 wt% graphene respectively.
Printing and Testing: An extrusion-based direct writing method was used, as described in previous works. [10,16,25] The substrates used involved YSZ and flexible Ribbon Ceramic (Alumina Ribbon Ceramic). Following the printing step, the print was allowed to dry under ambient conditions and subsequently sintered in a tube furnace at 800 °C for 30 min with a ramp-up of 5 °C min −1 under a forming gas environment (95% Nitrogen and 5% Hydrogen) for in-situ conversion of polydopamine to graphene and also enable the removal of excess residual organics on the surface. The four-point connections are made using Nichrome 80 wires (80% Nickel and 20% Cr) and fast-drying Ag paint. The electrical readings were characterized using Keithley 2450 Sourcemeter with a box furnace as the heating source. For wireless testing, Arduino IDE with ESP32 DEVKIT module was utilized. The module was connected to a Adafruit Pt100 RTD Temperature Sensor Amplifier and a resistor of 100 Ω was soldered acting as a reference. Four-point connections were utilized to connect the printed device to the amplifier, which interfaces with the ESP32 module and via a code written in Arduino IDE. ESP32 module interacts with the local server via the Wi-Fi connection.
Graphene Passivation: Graphene solution (0.5 wt% graphene flakes in water) was printed on top of the sintered copper conductor in varying thicknesses, which was controlled by the number of passes and the stand-off distance (the distance between the nozzle and the substrate). After drying, the passivated structure was again step-sintered in a tube furnace, first at 100 °C for 30 min followed by 600 °C for 30 min. The sintered conductor was utilized for characterization purposes.
Alumina Passivation: An atomic layer deposition setup was utilized for obtaining alumina layers in varying thicknesses. Trimethylaluminum (TMA) and water (H 2 O) were sequentially exposed to the sintered Cu-G conductor in durations of 20 s each, under the purge of nitrogen and a temperature of 200 °C. The number of cycles were determined by the desired thicknesses.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.